Molecular dynamics simulation of combustion reaction process and products of oxygen-containing functional groups in coal based on Machine Learning Potential

Abstract

Oxygen-containing functional groups are the main heat source of coal spontaneous combustion, but their complex reaction pathways and microscopic mechanisms are still unclear. In this study, the molecular dynamics simulation of the composite combustion reaction model system with four oxygen-containing functional groups was conducted using the Machine Learning (ML) potential force field. The results showed that the stability order of these four oxygen-containing functional groups in the combustion reaction system is as follows: ‒OH < ‒COOH < ‒C‒O < ‒C = O. These simulation findings align with those obtained from in-situ FTIR experiments, thereby validating the accuracy of the ML potential force field. Notably, ‒C‒O exhibits the highest tendency for CO₂ conversion (58.57 %); ‒COOH displays the highest tendency for H₂O conversion (45.00 %); and ‒OH demonstrates the highest tendency for CO conversion (59.17 %). The essence of the oxidative combustion reaction pathway involving oxygen-containing functional groups lies in the heat accumulation resulting from the oxidative dehydrogenation effect.